LONDON — Clinical trials have shown that lifestyle programs — which include diet, exercise, and behavioral coaching — can help people in danger of developing type 2 diabetes from tipping into a diagnosis of the condition. But there’s been a nagging question of whether such intensive regimens work in the real world.

A study published Wednesday backs up the idea that they can. Researchers behind the work relied on novel statistical approaches to analyze millions of records from England’s National Health Service and found that participants in the NHS’s Diabetes Prevention Program saw improvements in risk factors for type 2 diabetes, indicating that patients can benefit from such initiatives even outside the confines of a controlled experiment.

“It appears that if implemented at scale, such a fairly intensive lifestyle intervention still does have important health benefits for patients with prediabetes,” said Pascal Geldsetzer, an epidemiologist at Stanford University and the senior author of the paper, which was published in the journal Nature. “I think there is some skepticism among clinicians that providing lifestyle advice is often ineffective, but here we show that through a structured intensive program, it does have benefits.”

Past research, including the landmark Diabetes Prevention Program study in the U.S., has indicated that lifestyle interventions can reduce the chances of developing type 2 diabetes. But experts have debated how applicable those findings are. Trials might include individualized care with one-on-one sessions. People who enroll in such a study might be highly motivated to stick with the programs and have better access to transportation and information about improving their health.

In other words, would the results hold true for health systems and patients broadly? It’s a crucial public health question, as the rates of diabetes in countries including the U.S. and U.K. have climbed in recent decades, with fears of even higher rates in the future given the increased prevalence of obesity as well.

To answer the question, the research team turned to the NHS, which has data from millions of patients to study. It also has the Diabetes Prevention Program, a mix of education and support that includes weight loss, diet, and physical activity targets, with at least 13 group sessions over nine months.

Patients are only eligible for the program if they have blood-glucose levels — known as HbA1c — of at least 6%. (The figure is also measured in a unit called millimoles per mole, or mmol/mol.) An elevated HbA1c indicates that a person may have what’s called prediabetes, meaning they are essentially at risk for the condition.

For the researchers, that eligibility cutoff enabled them to build a study that was based on real-world data but that included “quasi-experimental” approaches. They could analyze data from patients on either side of that threshold and track how their glycemic control, as well as other risk factors, changed based on their eligibility for the program. Geldsetzer explained that they relied on statistical methods that are more commonly used in the social sciences than in epidemiology, but that are designed to determine whether an intervention has a causal effect, not just an association with an outcome.

And indeed, they found that people referred to the program had a drop in the HbA1c levels of 0.85 mmol/mol, based on one of the analyses.

“Our study not only demonstrates the potential of intensive lifestyle counseling for improving the health of patients with prediabetes in routine care but potentially also suggests a promising route for improving population health more broadly,” the authors wrote.

The study primarily focused on changes in blood-glucose levels, but also saw improvements in other risk factors such as BMI and weight. There were, however, no statistical effects on some markers, such as certain cholesterol levels, and on complications from diabetes or emergency hospitalizations for cardiovascular issues, though the authors noted they were studying a short follow-up period.

Mohammed Ali, the co-director of Emory University’s Global Diabetes Research Center, who was not involved in the research, praised the analysis and said the study was a valuable contribution to the conversation about the effectiveness of such programs. But he also said he wondered whether the statistical improvement seen in glucose levels was clinically meaningful.

“If I put on my clinical hat, and I am also a primary care doctor, and I see a lot of people with prediabetes, that’s something I think about,” said Ali, who added that other research has shown that lifestyle intervention programs don’t work for all types of prediabetes.

In the paper, the authors acknowledged that the “clinical significance” of such a reduction in HbA1c “is difficult to quantify at an individual level.” But they noted that there is a link between higher HbA1c and cardiovascular risk — suggesting that the reduction in glucose levels “is meaningful at the population level.”

The question of whether behavioral interventions work at a broad scale is important not just for the health of patients. Health systems also want to make sure they’re getting value out of such intensive efforts.

But having such a program is different than getting patients to participate, and helping them stick with it. The researchers reported that only 17.4% of eligible patients were referred to intensive lifestyle counseling.

In an editorial that was published along with the study Wednesday, Edward Gregg and Naomi Holman of Ireland’s RCSI University of Medicine and Health Sciences noted that the improvements in HbA1c were greater among people who participated in the program — on par with what past trials have shown — versus those who were referred but did not participate, and even more so than those of everyone who was eligible.

“This cascade of lowered risk — which was greatest in those who actively participated in the program — reflects the crux of the challenge for individual-level prevention approaches,” Gregg and Holman wrote. “The success of these strategies depends heavily on strong engagement and adherence to the program, and perhaps the identification of people who will respond well to lifestyle intervention in the first place. It is also a reminder that approaches targeted to the individual cannot go it alone: a combination of individual- and population-based approaches that address a wide variety of risk factors are needed to change the course of the type 2 diabetes epidemic.”